search

Dunedan / mbp-2016-linux

State of Linux on the MacBook Pro 2016 & 2017
2.1k stars 108 forks source link

MacBook Pro 15+: iBridge Support #110

Open aunali1 opened 5 years ago

aunali1 commented 5 years ago

Discussion and updates regarding iBridge related support. Parent thread is #71.

roadrunner2 commented 5 years ago

If folks could test the touchbar and light-sensor drivers on MacBookPro15,*, that would be very much appreciated. The driver is located at roadrunner2/macbook12-spi-driver (in the mbp15 branch). See the README at that link for build and install instructions.

Once installed, could you please test the following:

BernhardValenti commented 5 years ago

does it turn on and show ESC+brightness+multimedia keys?

yes

does pressing and releasing the fn key reliably toggle between the above and the function-keys?

no, just stays ESC+brightness+multimedia

does the above toggling still work after enabling alt mode?

no, just stays ESC+brightness+multimedia

does dimming work?

yes

does turning off the display work?

yes

you can also test automatic dimming/off by setting a short timeout;

yes

Ambient Light Sensor

couldn't get iio-sensor-proxy to work, so couldn't check

iggy commented 5 years ago

Can confirm the same results as above.

In addition, I tested the light sensor using the watch command and could see changes as I covered the sensor.

roadrunner2 commented 5 years ago

Thanks to both of you for testing!

Regarding the fn key issue: I think I found the problem and just pushed an update, so please re-pull. If it still doesn't work, then could you do two things:

Regarding the ALS: if the watch command works, then basically the driver is working, so that's good; iio-sensor-proxy is used by at least Gnome and KDE to get the sensor readings, i.e. if that's working then you should be able to configure things like automatic screen brightness. Also, if the ALS is working (irrespective of whether iio-sensor-proxy is installed or not) then the touchbar's brightness should automatically change as the sensor's reading changes.

BernhardValenti commented 5 years ago

@roadrunner2 just updated and tested. fn key still doesn't do anything. no errors in dmesg and the manual mode changes seem to work, not always but i'd say >95% of the time they work

roadrunner2 commented 5 years ago

@wedgeV Let's first look at the fn key issue then, as having that working will make testing the mode changes much easier. So somehow I must have have the input-device id matching wrong - could you dump me some info on the input devices on your system?

for inp in /sys/class/input/input* ; do cat $inp/name; grep . $inp/id/*; cat $inp/modalias; cat $inp/capabilities/key; echo ---; done
BernhardValenti commented 5 years ago 
/sys/class/input/input0/id/bustype:0019
/sys/class/input/input0/id/product:0005
/sys/class/input/input0/id/vendor:0000
/sys/class/input/input0/id/version:0000
input:b0019v0000p0005e0000-e0,5,kramlsfw0,
0
---
Power Button
/sys/class/input/input1/id/bustype:0019
/sys/class/input/input1/id/product:0001
/sys/class/input/input1/id/vendor:0000
/sys/class/input/input1/id/version:0000
input:b0019v0000p0001e0000-e0,1,k74,ramlsfw
10000000000000 0
---
HDA Intel PCH HDMI/DP,pcm=10
/sys/class/input/input10/id/bustype:0000
/sys/class/input/input10/id/product:0000
/sys/class/input/input10/id/vendor:0000
/sys/class/input/input10/id/version:0000
input:b0000v0000p0000e0000-e0,5,kramlsfw6,8,
0
---
Apple Headset
/sys/class/input/input11/id/bustype:0003
/sys/class/input/input11/id/product:5043
/sys/class/input/input11/id/vendor:05ac
/sys/class/input/input11/id/version:0101
input:b0003v05ACp5043e0101-e0,1,4,k72,73,A4,ram4,lsfw
1000000000 c000000000000 0
---
Apple Inc. Apple Internal Keyboard / Trackpad Keyboard
/sys/class/input/input12/id/bustype:0003
/sys/class/input/input12/id/product:027b
/sys/class/input/input12/id/vendor:05ac
/sys/class/input/input12/id/version:0101
input:b0003v05ACp027Be0101-e0,1,4,11,14,k71,72,73,74,75,77,79,7A,7B,7C,7D,7E,7F,80,81,82,83,84,85,86,87,88,89,8A,8C,8E,96,98,9E,9F,A1,A3,A4,A5,A6,AD,B0,B1,B2,B3,B4,B7,B8,B9,BA,BB,BC,BD,BE,BF,C0,C1,C2,F0,ram4,l0,1,2,3,4,sfw
1000000000007 ff9f207ac14057ff febeffdfffefffff fffffffffffffffe
---
Apple Inc. Apple Internal Keyboard / Trackpad Consumer Control
/sys/class/input/input13/id/bustype:0003
/sys/class/input/input13/id/product:027b
/sys/class/input/input13/id/vendor:05ac
/sys/class/input/input13/id/version:0101
input:b0003v05ACp027Be0101-e0,1,4,kA3,A4,A5,A8,D0,ram4,lsfw
10000 13800000000 0 0
---
Apple Inc. Apple Internal Keyboard / Trackpad Mouse
/sys/class/input/input16/id/bustype:0003
/sys/class/input/input16/id/product:027b
/sys/class/input/input16/id/vendor:05ac
/sys/class/input/input16/id/version:0101
input:b0003v05ACp027Be0101-e0,1,2,4,k110,111,112,r0,1,am4,lsfw
70000 0 0 0 0
---
Sleep Button
/sys/class/input/input2/id/bustype:0019
/sys/class/input/input2/id/product:0003
/sys/class/input/input2/id/vendor:0000
/sys/class/input/input2/id/version:0000
input:b0019v0000p0003e0000-e0,1,k8E,ramlsfw
4000 0 0
---
Apple Inc. iBridge Display
/sys/class/input/input20/id/bustype:0003
/sys/class/input/input20/id/product:8302
/sys/class/input/input20/id/vendor:05ac
/sys/class/input/input20/id/version:0101
input:b0003v05ACp8302e0101-e0,1,4,14,k71,72,73,78,A3,A4,A5,CC,E0,E1,E5,E6,F0,ram4,lsfw
1006300001000 3800000000 10e00000180001f f800000000000002
---
Power Button
/sys/class/input/input3/id/bustype:0019
/sys/class/input/input3/id/product:0001
/sys/class/input/input3/id/vendor:0000
/sys/class/input/input3/id/version:0000
input:b0019v0000p0001e0000-e0,1,k74,ramlsfw
10000000000000 0
---
PC Speaker
/sys/class/input/input4/id/bustype:0010
/sys/class/input/input4/id/product:0001
/sys/class/input/input4/id/vendor:001f
/sys/class/input/input4/id/version:0100
input:b0010v001Fp0001e0100-e0,12,kramls1,2,fw
0
---
Video Bus
/sys/class/input/input5/id/bustype:0019
/sys/class/input/input5/id/product:0006
/sys/class/input/input5/id/vendor:0000
/sys/class/input/input5/id/version:0000
input:b0019v0000p0006e0000-e0,1,kE0,E1,E3,F1,F2,F3,F4,F5,ramlsfw
3e000b00000000 0 0 0
---
HDA Intel PCH HDMI/DP,pcm=3
/sys/class/input/input6/id/bustype:0000
/sys/class/input/input6/id/product:0000
/sys/class/input/input6/id/vendor:0000
/sys/class/input/input6/id/version:0000
input:b0000v0000p0000e0000-e0,5,kramlsfw6,8,
0
---
HDA Intel PCH HDMI/DP,pcm=7
/sys/class/input/input7/id/bustype:0000
/sys/class/input/input7/id/product:0000
/sys/class/input/input7/id/vendor:0000
/sys/class/input/input7/id/version:0000
input:b0000v0000p0000e0000-e0,5,kramlsfw6,8,
0
---
HDA Intel PCH HDMI/DP,pcm=8
/sys/class/input/input8/id/bustype:0000
/sys/class/input/input8/id/product:0000
/sys/class/input/input8/id/vendor:0000
/sys/class/input/input8/id/version:0000
input:b0000v0000p0000e0000-e0,5,kramlsfw6,8,
0
---
HDA Intel PCH HDMI/DP,pcm=9
/sys/class/input/input9/id/bustype:0000
/sys/class/input/input9/id/product:0000
/sys/class/input/input9/id/vendor:0000
/sys/class/input/input9/id/version:0000
input:b0000v0000p0000e0000-e0,5,kramlsfw6,8,
0
---
roadrunner2 commented 5 years ago

Thanks, this clarifies things: the keybits are wonky, and that's because the relevant drivers don't know about the new usb id's, and hence aren't being loaded/attached. You can (temporarily) register the new id's as follows:

sudo modprobe hid-apple
echo 3 05ac 027b 4 | sudo tee /sys/bus/hid/drivers/apple/new_id

sudo modprobe bcm5974
echo 05ac 027b | sudo tee /sys/bus/usb/drivers/bcm5974/new_id

The touchbar driver may need to be reloaded after this, not sure.

ppaulweber commented 5 years ago

@wedgeV and @roadrunner2: I'm using the same patch https://github.com/Dunedan/mbp-2016-linux/issues/71#issuecomment-517208901 on my MacBookAir8,1 (2018 model with T2) and it works very well! If you have not compiled bcm5974 into the kernel, ensure that you load the bcm5974 module before the bce driver of @MCMrARM, then everything is setup correctly and you do not need to call a drivers new_id interface.

ppaulweber commented 5 years ago

@wedgeV and @roadrunner2: I just fixed the FN issue on my MacBookAir8,1 keyboard by applying the following kernel patch:

Linux 5.1.5 based patch to resolve keyboard FN key issues for MacBookAir8,1

```diff diff --git a/drivers/hid/hid-ids.h b/drivers/hid/hid-ids.h index adce58f24..9e46d7c16 100644 --- a/drivers/hid/hid-ids.h +++ b/drivers/hid/hid-ids.h @@ -170,6 +170,7 @@ #define USB_DEVICE_ID_APPLE_WELLSPRING9_ANSI 0x0272 #define USB_DEVICE_ID_APPLE_WELLSPRING9_ISO 0x0273 #define USB_DEVICE_ID_APPLE_WELLSPRING9_JIS 0x0274 +#define USB_DEVICE_ID_APPLE_MBA81 0x027a #define USB_DEVICE_ID_APPLE_FOUNTAIN_TP_ONLY 0x030a #define USB_DEVICE_ID_APPLE_GEYSER1_TP_ONLY 0x030b #define USB_DEVICE_ID_APPLE_IRCONTROL 0x8240 diff --git a/drivers/hid/hid-quirks.c b/drivers/hid/hid-quirks.c index 77ffba48c..a87bfb174 100644 --- a/drivers/hid/hid-quirks.c +++ b/drivers/hid/hid-quirks.c @@ -269,6 +269,7 @@ static const struct hid_device_id hid_have_special_driver[] = { { HID_USB_DEVICE(USB_VENDOR_ID_APPLE, USB_DEVICE_ID_APPLE_WELLSPRING9_ANSI) }, { HID_USB_DEVICE(USB_VENDOR_ID_APPLE, USB_DEVICE_ID_APPLE_WELLSPRING9_ISO) }, { HID_USB_DEVICE(USB_VENDOR_ID_APPLE, USB_DEVICE_ID_APPLE_WELLSPRING9_JIS) }, + { HID_USB_DEVICE(USB_VENDOR_ID_APPLE, USB_DEVICE_ID_APPLE_MBA81) }, { HID_BLUETOOTH_DEVICE(USB_VENDOR_ID_APPLE, USB_DEVICE_ID_APPLE_ALU_WIRELESS_2009_ANSI) }, { HID_BLUETOOTH_DEVICE(USB_VENDOR_ID_APPLE, USB_DEVICE_ID_APPLE_ALU_WIRELESS_2009_ISO) }, { HID_BLUETOOTH_DEVICE(USB_VENDOR_ID_APPLE, USB_DEVICE_ID_APPLE_ALU_WIRELESS_2009_JIS) }, @@ -916,6 +917,7 @@ static const struct hid_device_id hid_mouse_ignore_list[] = { { HID_USB_DEVICE(USB_VENDOR_ID_APPLE, USB_DEVICE_ID_APPLE_WELLSPRING9_ANSI) }, { HID_USB_DEVICE(USB_VENDOR_ID_APPLE, USB_DEVICE_ID_APPLE_WELLSPRING9_ISO) }, { HID_USB_DEVICE(USB_VENDOR_ID_APPLE, USB_DEVICE_ID_APPLE_WELLSPRING9_JIS) }, + { HID_USB_DEVICE(USB_VENDOR_ID_APPLE, USB_DEVICE_ID_APPLE_MBA81) }, { HID_USB_DEVICE(USB_VENDOR_ID_APPLE, USB_DEVICE_ID_APPLE_FOUNTAIN_TP_ONLY) }, { HID_USB_DEVICE(USB_VENDOR_ID_APPLE, USB_DEVICE_ID_APPLE_GEYSER1_TP_ONLY) }, { } ```

To apply this to an MacBookPro15,x please modify or add the 027b device. I've prepared another patch for you MBP guys, please let me know if it works out for you:

Linux 5.1.5 based patch to resolve keyboard FN issues for MacBookPro15,x

```diff diff --git a/drivers/hid/hid-ids.h b/drivers/hid/hid-ids.h index adce58f24..9e46d7c16 100644 --- a/drivers/hid/hid-ids.h +++ b/drivers/hid/hid-ids.h @@ -170,6 +170,7 @@ #define USB_DEVICE_ID_APPLE_WELLSPRING9_ANSI 0x0272 #define USB_DEVICE_ID_APPLE_WELLSPRING9_ISO 0x0273 #define USB_DEVICE_ID_APPLE_WELLSPRING9_JIS 0x0274 +#define USB_DEVICE_ID_APPLE_MBP15 0x027b #define USB_DEVICE_ID_APPLE_FOUNTAIN_TP_ONLY 0x030a #define USB_DEVICE_ID_APPLE_GEYSER1_TP_ONLY 0x030b #define USB_DEVICE_ID_APPLE_IRCONTROL 0x8240 diff --git a/drivers/hid/hid-quirks.c b/drivers/hid/hid-quirks.c index 77ffba48c..a87bfb174 100644 --- a/drivers/hid/hid-quirks.c +++ b/drivers/hid/hid-quirks.c @@ -269,6 +269,7 @@ static const struct hid_device_id hid_have_special_driver[] = { { HID_USB_DEVICE(USB_VENDOR_ID_APPLE, USB_DEVICE_ID_APPLE_WELLSPRING9_ANSI) }, { HID_USB_DEVICE(USB_VENDOR_ID_APPLE, USB_DEVICE_ID_APPLE_WELLSPRING9_ISO) }, { HID_USB_DEVICE(USB_VENDOR_ID_APPLE, USB_DEVICE_ID_APPLE_WELLSPRING9_JIS) }, + { HID_USB_DEVICE(USB_VENDOR_ID_APPLE, USB_DEVICE_ID_APPLE_MBP15) }, { HID_BLUETOOTH_DEVICE(USB_VENDOR_ID_APPLE, USB_DEVICE_ID_APPLE_ALU_WIRELESS_2009_ANSI) }, { HID_BLUETOOTH_DEVICE(USB_VENDOR_ID_APPLE, USB_DEVICE_ID_APPLE_ALU_WIRELESS_2009_ISO) }, { HID_BLUETOOTH_DEVICE(USB_VENDOR_ID_APPLE, USB_DEVICE_ID_APPLE_ALU_WIRELESS_2009_JIS) }, @@ -916,6 +917,7 @@ static const struct hid_device_id hid_mouse_ignore_list[] = { { HID_USB_DEVICE(USB_VENDOR_ID_APPLE, USB_DEVICE_ID_APPLE_WELLSPRING9_ANSI) }, { HID_USB_DEVICE(USB_VENDOR_ID_APPLE, USB_DEVICE_ID_APPLE_WELLSPRING9_ISO) }, { HID_USB_DEVICE(USB_VENDOR_ID_APPLE, USB_DEVICE_ID_APPLE_WELLSPRING9_JIS) }, + { HID_USB_DEVICE(USB_VENDOR_ID_APPLE, USB_DEVICE_ID_APPLE_MBP15) }, { HID_USB_DEVICE(USB_VENDOR_ID_APPLE, USB_DEVICE_ID_APPLE_FOUNTAIN_TP_ONLY) }, { HID_USB_DEVICE(USB_VENDOR_ID_APPLE, USB_DEVICE_ID_APPLE_GEYSER1_TP_ONLY) }, { } ```

Another remark to my previous comment: Due to this change I've updated my (arch) default loaded modules chain like this:

$ cat /etc/modules-load.d/bce.conf 
# Load bce.ko at boot
hid-apple
bcm5974
bce

And everything regarding bce-based keyboard and touchpad works just perfectly!

BernhardValenti commented 5 years ago

@ppaulweber cool, i'll try that! do you have two finger scrolling etc. working with the touchpad?

ppaulweber commented 5 years ago

@wedgeV you're welcome

do you have two finger scrolling etc. working with the touchpad?

Yes, everything works fine! You just really have to ensure that bce is loaded after the bcm5974 driver, which is the Linux-based Apple touchpad driver and when than the bce registers the new discovered USB devices, it assigns correctly the built-in touchpad with the hid-apple driver instead to the hid-generic which gives you the single click behavior.

BernhardValenti commented 5 years ago

@ppaulweber can you share your linux and linux-headers packages (or a git repo that i can build it from)? i'm also on arch, currently using the packages from the repo someone shared on the main issue

ppaulweber commented 5 years ago

@wedgeV OK, I'll prepare a GitHub repo containing a small config and build script, which applies the patches provided by @MCMrARM et al. and I'll try to merge the support for both platforms, namely MacBookAir8,1 and MacBookPro15,x.

ppaulweber commented 5 years ago

@wedgeV here you go https://github.com/ppaulweber/linux-mba, have fun and let me know if it works for you now

BernhardValenti commented 5 years ago

@ppaulweber thanks so much! works, i now have scrolling/right click working!

ppaulweber commented 5 years ago

@wedgeV perfect! I'm just putting together all the other guys (@MCMrARM et al.) out there have discovered in there heavy lifting. At this point a really bit thanks and nice work to you all :tada: :1st_place_medal:

GoncaloHit commented 5 years ago

I would like to test and help, but i have some questions, i have the MBP13.3 2019 i7 Touch bar, currently on windows, i wiped out osx, had to inject drivers in win boot setup, as for linux, does the latest kernel have the SSD drivers ? or do i need to build custom kernel with it, then flash it over an USB ? Thanks

mikeeq commented 5 years ago

Is keyboard backlight working for anybody?

MCMrARM commented 5 years ago

Good news, I got suspend working (sometimes, it's not reliable) and am working on pushing the code onto GitHub.

As for reliability, the whole OS sometimes just doesn't wake up, makes it really hard to debug.

MCMrARM commented 5 years ago

Also it might be worthwhile to try to reverse engineer some of the XPC comms, so that one could try to take over the shitty fan control as I think it's the powerchime service managing it.

MCMrARM commented 5 years ago

I have pushed all the suspend related code for VHCI under the suspend branch: https://github.com/MCMrARM/mbp2018-bridge-drv/tree/suspend Testing welcome and appreciated. Audio is currently unsupported on that branch.

aunali1 commented 5 years ago

@axxapy (Commenting here since its related) The reason why the touchbar did not work out of the box is most likely due to the missing apple-ib-* drivers which are in @roadrunner2's repo. In general, to get all the tested iBridge components working you would need:

  1. A patched kernel (aunali1/linux-mbp)
  2. @MCMrARM's bce driver from MCMrARM/mbp2018-bridge-drv
  3. @roadrunner2's apple-ib-* drivers from his mbp15 branch (roadrunner2/macbook12-spi-driver) NOTE: Currently his repo does not have HID IDs for MacbookPro15,1 and MacbookAir8,1 EDIT: As of https://github.com/roadrunner2/macbook12-spi-driver/commit/d05467440f16d17c2572c5fe3ca59f126758f00f keyboard/trackpad should be detected for all models.
roadrunner2 commented 5 years ago

Btw., sorry to ask again, but if anybody has the touchbar from my branch working as well as the patches to add the new id's to the bcm5974 and hid-apple drivers (so that in particular the FN key is properly handled), could you post your answers to the touchbar questions in the second comment of this issue? It would be much appreciated.

roadrunner2 commented 5 years ago

@aunali1 I'm confused about your patches: why is the applespi driver needed? I thought that models with a T2 chip had USB keyboards and trackpads. Are you saying that you have a machine that has a T2 chip but still an SPI keyboard/trackpad? If so, which model would that be then? And could you post the output from dmesg | grep applespi?

aunali1 commented 5 years ago

@roadrunner2 Disregard that, I was testing something for a non-T2 Macbook Pro.

roadrunner2 commented 5 years ago

@aunali1 Ok, thanks.

aunali1 commented 5 years ago

Btw., sorry to ask again, but if anybody has the touchbar from my branch working as well as the patches to add the new id's to the bcm5974 and hid-apple drivers (so that in particular the FN key is properly handled), could you post your answers to the touchbar questions in the second comment of this issue? It would be much appreciated.

Went ahead and did some testing, here are my results:

  • Touch Bar:

    • does it turn on and show ESC+brightness+multimedia keys?

Yes

  • does pressing and releasing the fn key reliably toggle between the above and the function-keys?

Yes

  • does the above toggling still work after enabling alt mode? 
    echo 1 | sudo tee /sys/bus/hid/devices/*/alt

Yes

  • does dimming work? 
    # dim
echo 0 | sudo tee /sys/bus/hid/devices/*/dim_timeout
# full brightness again
echo -2 | sudo tee /sys/bus/hid/devices/*/dim_timeout

Yes

  • does turning off the display work? 
    # off
echo 0 | sudo tee /sys/bus/hid/devices/*/idle_timeout
# on again
echo 300 | sudo tee /sys/bus/hid/devices/*/idle_timeout

Yes

  • you can also test automatic dimming/off by setting a short timeout; e.g.

    echo 10 | sudo tee /sys/bus/hid/devices/*/idle_timeout

    The Touch Bar should then dim after about 7 seconds of inactivity, and then turn off after another 3 seconds; touching the Touch Bar or trackpad or hitting any key should wake it to full brightness again.

Yep, works.

  • Ambient Light Sensor: The light sensor is located at the top of the lid next to the camera. The easiest way to test is by installing the iio-sensor-proxy package (should be available on most distros) and then running the monitor-sensor command - it should say that there is a light sensor, and changing the brightness (e.g. by covering the sensor with your hand) should result in light-changed events. Alternatively, the light values can also be watched with some simple like: 
    watch -n 1 cat /sys/bus/iio/devices/iio:device0/in_illuminance_input

Works like a charm.

roadrunner2 commented 5 years ago

@aunali1 Thanks for testing and reporting back! This is good news - now I can integrate the changes for these new models back into the original and get back to upstreaming the thing...

mikeeq commented 5 years ago

@roadrunner2 any clue how to make keyboard backlight working? Due to lack of apple smc driver keyboard backlight device is not initialized, should it be done by your spi driver or in @MCMrARM bce one?

➜ ls -l /sys/class/leds                  
total 0
lrwxrwxrwx. 1 root root 0 Aug 28 11:11 input15::capslock -> ../../devices/pci0000:00/0000:00:1d.4/0000:02:00.1/bce/bce/bce-vhci/usb7/7-5/7-5:1.1/0003:05AC:027B.0004/input/input15/input15::capslock
lrwxrwxrwx. 1 root root 0 Aug 28 11:11 input15::compose -> ../../devices/pci0000:00/0000:00:1d.4/0000:02:00.1/bce/bce/bce-vhci/usb7/7-5/7-5:1.1/0003:05AC:027B.0004/input/input15/input15::compose
lrwxrwxrwx. 1 root root 0 Aug 28 11:11 input15::kana -> ../../devices/pci0000:00/0000:00:1d.4/0000:02:00.1/bce/bce/bce-vhci/usb7/7-5/7-5:1.1/0003:05AC:027B.0004/input/input15/input15::kana
lrwxrwxrwx. 1 root root 0 Aug 28 11:11 input15::numlock -> ../../devices/pci0000:00/0000:00:1d.4/0000:02:00.1/bce/bce/bce-vhci/usb7/7-5/7-5:1.1/0003:05AC:027B.0004/input/input15/input15::numlock
lrwxrwxrwx. 1 root root 0 Aug 28 11:11 input15::scrolllock -> ../../devices/pci0000:00/0000:00:1d.4/0000:02:00.1/bce/bce/bce-vhci/usb7/7-5/7-5:1.1/0003:05AC:027B.0004/input/input15/input15::scrolllock

And yeah I can confirm that your driver is working like a charm on my Macbook 15,2, all tb modes, diming etc. is working as mentioned above. I had some kernel panics caused by apple-ib-tb driver during boot time but i haven't got a chance to capture it :(

MCMrARM commented 5 years ago

well my driver is not providing any other functionality that bringing the USB hub up and providing audio support, so it definitely should not be handled by mine

MCMrARM commented 5 years ago

HID reports for setting up the backlight: https://mrarm.io/u/Paste%202019-08-28%2013-51-35.txt

mikeeq commented 5 years ago

@roadrunner2 I've tried to test your latest changes: https://github.com/roadrunner2/macbook12-spi-driver/tree/d05467440f16d17c2572c5fe3ca59f126758f00f

but apple-ib-tb.ko module is crashing during load time:

[   91.756773] bce_vhci_urb_dequeue 81
[   91.764955] input: Apple Inc. iBridge Display as /devices/pci0000:00/0000:00:1d.4/0000:02:00.1/bce/bce/bce-vhci/usb7/7-6/7-6:1.0/0003:05AC:8302.0008/input/input16
[   91.816868] apple-ib-touchbar 0003:05AC:8302.0008: input: USB HID v1.01 Keyboard [Apple Inc. iBridge Display] on usb-bce-vhci-6/input0
[   91.817082] apple-ib-touchbar 0003:05AC:8102.0009: : USB HID v1.01 Device [Apple Inc. iBridge DFR brightness] on usb-bce-vhci-7/input0
[   91.869650] BUG: unable to handle kernel NULL pointer dereference at 0000000000000000
[   91.869653] #PF error: [normal kernel read fault]
[   91.869654] PGD 0 P4D 0 
[   91.869656] Oops: 0000 [#1] SMP PTI
[   91.869658] CPU: 4 PID: 9884 Comm: insmod Tainted: G           OE     5.1.19-300.wifi.patch.fc30.x86_64 #1
[   91.869659] Hardware name: Apple Inc. MacBookPro15,2/Mac-827FB448E656EC26, BIOS 220.270.99.0.0 (iBridge: 16.16.6568.0.0,0) 06/25/2019
[   91.869662] RIP: 0010:appletb_match_internal_device+0x33/0x70 [apple_ib_tb]
[   91.869663] Code: 1c 48 8d 86 18 02 00 00 41 b8 01 00 00 00 75 29 eb 2f 48 8b 50 58 48 85 d2 74 1b 48 8b 32 48 c7 c7 34 75 dc c0 b9 0b 00 00 00 <f3> a6 0f 97 c2 80 da 00 84 d2 74 0f 48 8b 00 48 85 c0 75 d4 45 31
[   91.869664] RSP: 0018:ffffa8d18290ba50 EFLAGS: 00010282
[   91.869665] RAX: ffff999793e67a18 RBX: ffff99977fd98e58 RCX: 000000000000000b
[   91.869666] RDX: ffffffffb0ecb080 RSI: 0000000000000000 RDI: ffffffffc0dc7534
[   91.869666] RBP: ffff999793e67800 R08: 0000000000000001 R09: 0000000000000002
[   91.869667] R10: 0000000000000000 R11: 0000000000000001 R12: ffffffffc0dc78d0
[   91.869667] R13: ffff99977fd98ec0 R14: ffff99977fd98e58 R15: ffffffffc0dc9040
[   91.869668] FS:  00007fee958bb2c0(0000) GS:ffff99979db00000(0000) knlGS:0000000000000000
[   91.869669] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[   91.869669] CR2: 0000000000000000 CR3: 00000004345c0004 CR4: 00000000003606e0
[   91.869670] Call Trace:
[   91.869673]  ? input_attach_handler+0x38/0xa0
[   91.869675]  ? input_register_handler+0x8f/0xe0
[   91.869676]  ? appletb_probe+0x23f/0x3c0 [apple_ib_tb]
[   91.869678]  ? hid_device_probe+0x109/0x170
[   91.869681]  ? really_probe+0xf9/0x3a0
[   91.869682]  ? driver_probe_device+0xb6/0x100
[   91.869683]  ? driver_allows_async_probing+0x50/0x50
[   91.869684]  ? bus_for_each_drv+0x7b/0xc0
[   91.869686]  ? __device_attach+0xe4/0x160
[   91.869687]  ? __hid_register_driver+0x80/0x80
[   91.869688]  ? bus_rescan_devices_helper+0x3a/0x80
[   91.869689]  ? device_reprobe+0x18/0x30
[   91.869690]  ? bus_for_each_dev+0x78/0xc0
[   91.869691]  ? hid_destroy_device+0x60/0x60
[   91.869692]  ? __hid_bus_driver_added+0x28/0x30
[   91.869693]  ? bus_for_each_drv+0x7b/0xc0
[   91.869694]  ? __hid_register_driver+0x70/0x80
[   91.869695]  ? appletb_init+0xa8/0x1000 [apple_ib_tb]
[   91.869697]  ? 0xffffffffc0dcc000
[   91.869698]  ? do_one_initcall+0x46/0x1c4
[   91.869700]  ? free_unref_page_commit+0x95/0x110
[   91.869702]  ? _cond_resched+0x15/0x30
[   91.869703]  ? kmem_cache_alloc_trace+0x154/0x1c0
[   91.869705]  ? do_init_module+0x23/0x210
[   91.869706]  ? do_init_module+0x5c/0x210
[   91.869706]  ? load_module+0x23de/0x2910
[   91.869709]  ? ima_post_read_file+0xf0/0x100
[   91.869710]  ? __do_sys_finit_module+0xaa/0x110
[   91.869711]  ? __do_sys_finit_module+0xaa/0x110
[   91.869713]  ? do_syscall_64+0x5b/0x170
[   91.869714]  ? entry_SYSCALL_64_after_hwframe+0x44/0xa9
[   91.869715] Modules linked in: apple_ib_tb(OE+) apple_ibridge(OE) nls_utf8 isofs rfcomm nf_conntrack_netlink xt_addrtype br_netfilter xt_CHECKSUM ipt_MASQUERADE tun bridge stp llc nf_conntrack_netbios_ns nf_conntrack_broadcast xt_CT ip6t_rpfilter ip6t_REJECT nf_reject_ipv6 ipt_REJECT nf_reject_ipv4 xt_conntrack ebtable_nat ip6table_nat ip6table_mangle ip6table_raw ip6table_security iptable_nat nf_nat iptable_mangle iptable_raw iptable_security nf_conntrack nf_defrag_ipv6 nf_defrag_ipv4 libcrc32c ip_set nfnetlink overlay ebtable_filter ebtables ip6table_filter ip6_tables iptable_filter ccm des_generic arc4 bnep cmac sha512_ssse3 sha512_generic md4 sunrpc vfat fat uvcvideo videobuf2_vmalloc videobuf2_memops videobuf2_v4l2 videobuf2_common videodev iTCO_wdt mei_hdcp iTCO_vendor_support joydev media snd_soc_skl intel_rapl x86_pkg_temp_thermal intel_powerclamp snd_soc_hdac_hda snd_hda_ext_core snd_soc_skl_ipc coretemp snd_soc_sst_ipc snd_soc_sst_dsp kvm_intel snd_soc_acpi_intel_match
[   91.869733]  snd_soc_acpi snd_soc_core kvm brcmfmac snd_compress snd_hda_codec_hdmi ac97_bus snd_pcm_dmaengine brcmutil irqbypass intel_cstate snd_hda_intel intel_uncore cfg80211 snd_hda_codec intel_rapl_perf snd_hda_core idma64 hci_uart snd_hwdep i2c_i801 btqca btrtl intel_lpss_pci btbcm intel_lpss thunderbolt btintel bluetooth mmc_core mei_me sbs mei ecdh_generic intel_pch_thermal rfkill sbshc acpi_tad pcc_cpufreq apple_bl ip_tables dm_crypt cdc_mbim cdc_wdm nvme cdc_ncm usbnet nvme_core mii i915 uas usb_storage i2c_algo_bit drm_kms_helper crct10dif_pclmul crc32_pclmul crc32c_intel drm ghash_clmulni_intel video pkcs8_key_parser fuse bce(OE) snd_pcm snd_seq snd_timer snd_seq_device snd soundcore bcm5974 hid_apple
[   91.869750] CR2: 0000000000000000
[   91.869751] ---[ end trace a79c0a5ec2a52066 ]---
[   91.869753] RIP: 0010:appletb_match_internal_device+0x33/0x70 [apple_ib_tb]
[   91.869754] Code: 1c 48 8d 86 18 02 00 00 41 b8 01 00 00 00 75 29 eb 2f 48 8b 50 58 48 85 d2 74 1b 48 8b 32 48 c7 c7 34 75 dc c0 b9 0b 00 00 00 <f3> a6 0f 97 c2 80 da 00 84 d2 74 0f 48 8b 00 48 85 c0 75 d4 45 31
[   91.869755] RSP: 0018:ffffa8d18290ba50 EFLAGS: 00010282
[   91.869755] RAX: ffff999793e67a18 RBX: ffff99977fd98e58 RCX: 000000000000000b
[   91.869756] RDX: ffffffffb0ecb080 RSI: 0000000000000000 RDI: ffffffffc0dc7534
[   91.869756] RBP: ffff999793e67800 R08: 0000000000000001 R09: 0000000000000002
[   91.869757] R10: 0000000000000000 R11: 0000000000000001 R12: ffffffffc0dc78d0
[   91.869757] R13: ffff99977fd98ec0 R14: ffff99977fd98e58 R15: ffffffffc0dc9040
[   91.869758] FS:  00007fee958bb2c0(0000) GS:ffff99979db00000(0000) knlGS:0000000000000000
[   91.869759] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[   91.869759] CR2: 0000000000000000 CR3: 00000004345c0004 CR4: 00000000003606e0
roadrunner2 commented 5 years ago

@mikeeq Thanks for the report - just pushed a fix.

Btw., what model do you have? I'm trying to understand why it crashed for you but not for others.

mikeeq commented 5 years ago

@roadrunner2 - Macbook Pro 15,2 13" 2019 i5 TouchBar Z0WQ000AR MV972ZE/A/R1

I can confirm that patched version is working ;) Will you have a minute to implement keyboard backlight to your ibridge driver?

MCMrARM commented 5 years ago

I'm working on patching the hid_apple driver to support backlight

MCMrARM commented 5 years ago

https://github.com/MCMrARM/mbp2018-etc/tree/master/apple-hid You can patch the kernel using the provided patch or compile the HID as a module.

mikeeq commented 5 years ago

@MCMrARM are you able to change keyboard backlight brightness by pressing buttons on touchbar?

but i can confirm that it's working for me - thanks Man. Good job, once again! 

➜ ls -l /sys/class/leds/apple::kbd_backlight/
total 0
-rw-r--r--. 1 root root 4096 Aug 29 17:28 brightness
lrwxrwxrwx. 1 root root    0 Aug 29 17:28 device -> ../../../0003:05AC:027B.0004
-r--r--r--. 1 root root 4096 Aug 29 17:28 max_brightness
drwxr-xr-x. 2 root root    0 Aug 29 17:28 power
lrwxrwxrwx. 1 root root    0 Aug 29 17:28 subsystem -> ../../../../../../../../../../../../../class/leds
-rw-r--r--. 1 root root 4096 Aug 29 17:28 trigger
-rw-r--r--. 1 root root 4096 Aug 29 17:26 uevent

One thing @roadrunner2 with the latest version of apple-ib-tb.ko fn keys are not showing up when pressing FN key (switching from media keys to fn is not working on default fnmode = 1). It's working ok on module built from this commit - https://github.com/roadrunner2/macbook12-spi-driver/tree/3f01cdb5035c4bc6aceeb2e835ded30699b3c06e

MCMrARM commented 5 years ago

@mikeeq no, I don't load the touchbar drivers during testing.

But are they supposed to just work? I thought it's supposed to be handled by userspace.

aunali1 commented 5 years ago

@mikeeq Does the following command return anything?

for inp in /sys/class/input/input* ; do cat $inp/name; grep . $inp/id/*; cat $inp/modalias; echo ---; done | grep "Internal Keyboard"
mikeeq commented 5 years ago

@aunali1 Yes it does.

➜ for inp in /sys/class/input/input* ; do cat $inp/name; grep . $inp/id/*; cat $inp/modalias; echo ---; done | grep "Internal Keyboard"

Apple Inc. Apple Internal Keyboard / Trackpad

And yeah @MCMrARM you were right - keyboard backlight brightness control after gnome restart is working for me both from GNOME settings gui and touchbar ;)

roadrunner2 commented 5 years ago

@MCMrARM nice work on the keyboard backlight driver! Sorry, didn't have time to look at it myself. Looking over the report info you provided, I noticed that the descriptor of the report to set the brightness is not present in the list of report descriptors for that interface, only the descriptor for the brightness config :frowning_face: .

A couple small things: the keyboard backlight doesn't turn unless the value is at least 32 (see backlight-min-on in the config) - in the applespi driver I therefore took the approach of scaling the given brightness values appropriately so that it turns on at value 1 (i.e. set-value = input-value / (max - min) * max). Something you might want to consider doing too. Also, I noticed that the max-brightness is 512, but in the kernel LED_FULL is defined as 255 - have you tested whether whether things like Gnome will actually scale their range between 0 to max, or assume max is always (at most) 255 (this can be tested by setting the keyboard brightness to max via the appropriate touchbar keys, and then looking at the actual brightness value in sysfs)? I.e. I'm wondering if it's necessary to set cdev.max_brightness to LED_FULL instead and scale the input values up to 0 - 512.

@mikeeq no, I don't load the touchbar drivers during testing.

But are they supposed to just work? I thought it's supposed to be handled by userspace.

The keyboard brightness keys are indeed just supposed to work; however, due to a bug in upower and gsd you usually have to restart upower (sudo systemctl restart upower) after (re)loading the driver providing the keyboard brightness controls (hid-apple in your case, applespi in models with T1 chip).

roadrunner2 commented 5 years ago

@mikeeq

One thing @roadrunner2 with the latest version of apple-ib-tb.ko fn keys are not showing up when pressing FN key (switching from media keys to fn is not working on default fnmode = 1). It's working ok on module built from this commit - https://github.com/roadrunner2/macbook12-spi-driver/tree/3f01cdb5035c4bc6aceeb2e835ded30699b3c06e

Hmm, that's odd. Could you run a slightly modified version of what you ran at @aunali1 request above?

for inp in /sys/class/input/input* ; do cat $inp/name; grep . $inp/id/*; cat $inp/capabilities/key; echo ---; done | grep -A5 "Internal Keyboard"

I want to doubly confirm the FN key is being properly "exported" by the keyboard driver.

Next, please apply the following patch to your apple-ib-tb.c:

--- a/apple-ib-tb.c
+++ b/apple-ib-tb.c
@@ -1130,6 +1130,10 @@ static bool appletb_match_internal_device(struct input_handler *handler,
                       strcmp(dev->type->name, "usb_device")))
                dev = dev->parent;

+       pr_info("match-internal-dev: bustype=%x vendor=%x product=%x dev=%p usb-prod='%s'\n",
+               inp_dev->id.bustype, inp_dev->id.vendor, inp_dev->id.product,
+               dev, dev ? to_usb_device(dev)->product : "-");
+
        /*
         * Apple labels all their internal keyboards and trackpads as such,
         * instead of maintaining an ever expanding list of product-id's we

And then show me the output of dmesg | grep match-internal after reloading the module. TIA.

mikeeq commented 5 years ago

@roadrunner2 

➜ for inp in /sys/class/input/input* ; do cat $inp/name; grep . $inp/id/*; cat $inp/capabilities/key; echo ---; done | grep -A5 "Internal Keyboard"

Apple Inc. Apple Internal Keyboard / Trackpad
/sys/class/input/input7/id/bustype:0003
/sys/class/input/input7/id/product:027b
/sys/class/input/input7/id/vendor:05ac
/sys/class/input/input7/id/version:0101
10000 0 0 0 1007b00011007 ff9f217ac14057ff ffbeffdfffefffff fffffffffffffffe

➜ dmesg | grep match-internal
[ 2902.630385] match-internal-dev: bustype=3 vendor=5ac product=27b dev=00000000528c5ebf usb-prod='Apple Internal Keyboard / Trackpad'
[ 2902.630393] match-internal-dev: bustype=3 vendor=5ac product=27b dev=00000000528c5ebf usb-prod='Apple Internal Keyboard / Trackpad'

As I said before FN switch is working properly on the previous version of the driver. On the new one besides switching everything works perfectly fine ;)

roadrunner2 commented 5 years ago

@mikeeq Thanks. Aaargh: I managed to get the string-search test inverted! :man_facepalming: Just pushed the fix.

mikeeq commented 5 years ago

Yep, it's working fine now - GJ @roadrunner2 ! Cheers ;)

MCMrARM commented 5 years ago

@roadrunner2 ah right, I didn't think of that. As for max_brightness it afaik should be fine, but I could verify it a bit later. Before upstreaming this patch I'll probably also want to get rid of the hardcoded constants in some places.

ppaulweber commented 5 years ago

@MCMrARM thanks for the keyboard backlight patch, I was able to integrate this one along with the changes for the MacBookAir8,x and it just works, well done :tada:

MCMrARM commented 5 years ago

I have been working on some more of the Touch ID integration stuff.... the protocol is absolutely insane, and gigantic: I've decided to list all the commands I can find and try to name them (the biometrickitd daemon handling this is written in Objective C... free function names) https://github.com/MCMrARM/bkitd/blob/master/BkKernelCommand.h#L10 this is 61 commands.... Anyways, some basic stuff work (like loading the calibration info), however trying to do enroll fails with -3, whatever that means honestly.

imbushuo commented 5 years ago

@MCMrARM I think SEP (the other PCIe device) is absolutely involved in any enrollment/verification according to my rough inspection in IDA.